变压器在线监测与故障诊断系统的研究

发布时间：2018-02-25 09:08

本文关键词： 电力变压器油中溶解气体分析故障诊断光声光谱　出处：《武汉理工大学》2014年硕士论文　论文类型：学位论文

【摘要】：电力变压器安全可靠地运行是整个供电系统稳定的基础。若设备出现故障，通常会引起局部的或全部的系统设备停止运行，造成严重的断电事故，并会带来巨大的经济损失。因此为了防止变压器出现故障，采用对变压器状态进行在线监测，来及早发现变压器的潜伏性故障，并及时采取措施，对变压器进行维修或更换，，便能够有效避免因重大事故的发生而造成不可预计的严重后果。油中溶解气体分析（DGA）技术因其不需要停电就可以对运行中的变压器的状态进行检测，而且也不会受到外界环境的干扰，是变压器状态监测中最为广泛使用的一种技术。它通过对采样的变压器油中所含气体的组分与浓度进行分析，来判断该变压器的运行状态。本论文以油中溶解气体分析技术为基础，对变压器在线监测与故障诊断系统进行研究。论文从油中溶解气体在线监测装置、故障诊断算法以及变压器在线监测管理系统三大模块分别进行讨论，构成一个完整的变压器状态监测与故障诊断系统：（1）油中溶解气体在线监测装置主要由油气分离和气体检测两部分组成。通过对比分析，确定了采用动态顶空脱气的方式进行油气分离，并使用光声光谱法对分离出的混合气体进行检测。实验数据显示，该装置的精确度高、稳定性好。（2）介绍了几种通过对气体浓度进行比值计算来确定变压器故障性质的传统算法。然后研究了一种动态加权的模糊聚类算法对变压器故障进行判断，该算法可以防止传统算法中因编码缺失或边界问题而造成诊断结果的不准确。（3）最后，设计了一款电力变压器在线监测管理系统，采用B/S架构，不再需要复杂繁琐的软件的安装与更新，可直接通过互联网实时获取变压器的运行状态数据。并且可以对运行中的变压器进行在线的故障诊断，通过可视化的诊断方式来确定变压器的状态信息。
[Abstract]:The safe and reliable operation of power transformers is the basis of the stability of the whole power supply system. Therefore, in order to prevent transformer failure, on-line monitoring of transformer condition is adopted to detect the latent fault of transformer as soon as possible, and timely measures are taken to repair or replace the transformer. Can effectively avoid the occurrence of serious accidents and unpredictable serious consequences. The dissolved Gas Analysis in Oil (DGA) technology can detect the status of transformers in operation without power outages, and will not be disturbed by the external environment. It is the most widely used technology in transformer condition monitoring. It can judge the operation state of transformer by analyzing the composition and concentration of gas in the sampled transformer oil. Based on the analysis technology of dissolved gas in oil, the on-line monitoring and fault diagnosis system of transformer is studied in this paper. The three modules of fault diagnosis algorithm and transformer on-line monitoring and management system are discussed, which constitute a complete transformer condition monitoring and fault diagnosis system. 1) On-line monitoring device for dissolved gas in oil is mainly composed of two parts: oil and gas separation and gas detection. Through comparison and analysis, it is determined that dynamic headspace degassing is used to separate oil and gas. The experimental data show that the device has high accuracy and good stability. This paper introduces several traditional algorithms to determine the fault properties of transformer by calculating the ratio of gas concentration, and then studies a dynamic weighted fuzzy clustering algorithm to judge transformer fault. This algorithm can prevent the inaccuracy of the diagnosis result caused by the missing coding or boundary problem in the traditional algorithm. Finally, a power transformer on-line monitoring and management system is designed, which adopts the B / S structure and no longer needs complicated software installation and update. The running state data of transformers can be obtained directly through the Internet, and the on-line fault diagnosis of transformers in operation can be carried out, and the state information of transformers can be determined by visual diagnosis.
【学位授予单位】：武汉理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM407

【参考文献】